Policies and practices for managing SBC LTER research protocols

Guidelines and templates

(requires login)

Definitions of 'Protocol' and 'Method'

Technically, the two terms are different: Protocols are 'prescriptive' and methods are 'descriptive'. That is, a protocol is a set of instructions for how to do a task - like a recipe - and a method describes how you actually did the task. Generally, the protocol gets written before (e.g., so it can be taken into the field), and the methods after (methods are written up for a paper). So your methods could be as simple as "we followed the x protocol." A protocol might say "filter 250 ml water", but your methods might say "if the filter clogged, we filtered only 100 ml."

In reality, what we call a 'protocol document' often includes both protocol and methods, especially when a single document is updated with protocol changes. For now this is unavoidable.

We publish protocols (or methods) with our datasets. It helps to assure that even our old data will be intelligible and interpretable far in the future. So all our protocol docs are archived as PDFs so they can be accessed with commonly used free software. It may seem like a chore, but the LTER is considered by NSF and many others to be an example of good data management practices, so we want to keep our reputation shiny. The policies and practices for SBC LTER datasets are outlined below.

There are 2 kinds of protocols: those which are SBC-originated (endogenous) and those which are lab-originated (or exogenous). The two types differ in the style of maintenance, and in how they are named and attached to datasets. We do not require collaborating researchers to create a protocol document specifically for SBC.

  • The SBC-originated protocols (endogenous) were written for the SBC LTER project, although they may be modifications of those used by researchers for their other projects. SBC-originated protocols are maintained over time in a single document and changes recorded there. Examples include most of the reef group's protocols for community surveys.
  • Lab-originated protocols are exogenous to SBC, i.e., they were developed by individual reaserchers in their own labs, and adopted by SBC mainly through collaboration. So these protocols are usually already standard practice in a researcher's laboratory. Most of the non-reef protocols fall into this category, e.g., the DOC filtration protocols from the Carlson lab, mooring data collection and processing from the Washburn lab, or stream gauge and chemistry sampling from the Melack laboratory.

Endogenous protocols are generally easier to maintain and keep track of, mainly because these are handled by staffers dedicated to SBC. Our use of exogenous protocols is harder to track, because SBC does not control when, why or how contributing laboratories manage their work; we rely on their expertise. All protocols are associated with datasets, so tracking protocols is more complicated when exogenous protocols are used in SBC long-term time-series. We have developed some recommendations for those tasked with handling datasets and their associated protocols (or methods). As of this writing (mid 2011), our use of protocols has not always followed all 'recommended practices'.

  1. File Location:
    Put the final versions of protocol docs in your group's /Final/Protocols/. The IM will retrieve the PDF version from there. If you don't have the software to create a PDF, a word-doc or text file is fine, and the IM can build the PDF. /Final/ is the directory where other SBC LTER people will go to find the material you intend to be shared (the /Working/ folder is for work-in-progress).
    It is recommended to archive the older versions in Final, e.g., under Ocean/Final/Protocols/ARCHIVE.
  2. File naming and versions:
    Exogenous protocol document will include:
    1. the name of the lab where it originated and
    2. a publication date.
    This way, we can preserve our old protocols for linking to datasets. So your file "CHN_filtration.pdf" will become "Brzezinski_20090601_CHN_filtration.pdf". when it's archived. This also gives us information on the history of the protocol, e.g., if the CHN analysis were to move to a different lab. Many of the protocols already linked to datasets do not have dates or names as part of their file names, but new ones will. A good example is the cruise datasets, which used 3 or 4 different primary production protocols, and are dated.
    Note: We have not yet established naming or dating conventions for SBC-originated protocols. However, it would make sense to prefix these files names with "SBC_LTER_", or the investigator's name and a date, since it is likely that these will eventually be shared beyond our group.
  3. Recording changes in time series data:
    Since protocols change over time, we should be stating in the dataset when we switch. So when a protocol changes, please let the IM know. The IM should also set up scripts to audit additions or changes to the Final/Protocol directory.
    At some point in the future, we should add this additonal material to time series. This is probably tedious, but hopefully this is recorded in the sample logs. When the dates are known, the IM can add text to the dataset, as in "this protocol was followed 2002-2006".

Note on difference between 'Protocol' and 'Method'